I've just noticed in source file webalizer.c :
/* month names used for parsing logfile (shouldn't be lang specific)
*/
char *log_month[12]={ "jan", "feb", "mar", "apr", "may", "jun",
"jul", "aug", "sep", "oct", "nov", "dec"};
and I realized in Italian it's ok for FEB, MAR, APR but NOT for MAY!!!
In fact, from error messages I get
Errore: Tralascio il record (data errata): [21/lug/2015:07:35:56 -0000]
[1]
Errore: Tralascio il record (data errata): [21/lug/2015:07:35:56 -0000]
[2]
Errore: Tralascio il record (data errata): [21/lug/2015:07:35:56 -0000]
[3]
Errore: Tralascio il record (data errata): [21/lug/2015:07:35:56 -0000]
[4]
Errore: Tralascio il record (data errata): [21/lug/2015:07:35:56 -0000]
[5]
can you see 'lug' instead of 'jul'....
So I think it can be an explanation for what I get now...
For Italy we should have:
char *log_month[12]={ "gen", "feb", "mar", "apr", "mag", "giu",
"lug", "ago", "set", "ott", "nov", "dic"};
but this can be a solution only for Italy and not for any other country
with a specific language other than english....
So:
1. Is it possible to chain the char *log_month[12]=... definition with
an external, locale-dependant, array?
or
2. is it possible to use another date-conversion function to get a
human-readable format from a unix timestamp ? (I remember... it's coming
from a standard squid log file)
Now, in parse.c i see:
/* format date/time field */
strftime(log_rec.datetime,sizeof(log_rec.datetime), "[%d/%b/%Y:%H:%M:
%S -0000]",localtime(&i));
and, as we can see form the man page,
http://man7.org/linux/man-pages/man3/strftime.3.html
%b The abbreviated month name according to the current locale.
Do You think we can found a workaround for this?
Why don't use
%m The month as a decimal number (range 01 to 12)
?
Thanx for the collaboration
Luca
